You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by michael8 <mi...@saracatech.com> on 2009/09/24 23:40:05 UTC

Looking for suggestion of WordDelimiter filter config and 'ALMA awards'

Hi,

I have this situation that I believe is very common but was curious if
anyone knows the right way to go about solving it.  

I have a document with 'ALMA awards' in it.  However, when user searches for
'aLMA awards', it ends up with no results found.  However, when I search for
'alma awards' or 'ALMA awards', the right results came back as expected.  

I immediately went to solr/admin/analysis to see what is going on with
indexing of 'ALMA awards' and query parsing of 'aLMA awards', and looks like
WordDelimiter is the one causing the mismatched.  WordDelimiter, with
splitOnCaseChange=1, will turn my search query 'aLMA awards' into 'a' and
'LMA' and 'awards', which is exactly what splitOnCaseChange does.  In this
type of situation, is there a proper way to handle such a situation whereby
the user simply got the case wrong for the 1st letter, or maybe n letters? 
I like the benefits that WordDelimiter filter w/ splitOnCaseChange provides
me, but I am not sure what is the proper way to solve this situation without
compromising on the other benefits this filter provides.  I also tried
preserveOriginal=1, hoping that aLMA will be preserved and later on became
all lowercase alma via another filter, but with no luck.

P.S.: I am basically using the standard config for 'text' fieldtype for my
default search field. (solr 1.3)

Thanks,
Michael
-- 
View this message in context: http://www.nabble.com/Looking-for-suggestion-of-WordDelimiter-filter-config-and-%27ALMA-awards%27-tp25591381p25591381.html
Sent from the Solr - User mailing list archive at Nabble.com.